soil organic carbon partitioning and Δ14c variation in ... · soil organic carbon partitioning and...
TRANSCRIPT
Soil organic carbon partitioning and D14C variationin desert and conifer ecosystems of southern Arizona
Rebecca A. Lybrand . Katherine Heckman . Craig Rasmussen
Received: 2 September 2016 / Accepted: 29 June 2017 / Published online: 13 July 2017
� Springer International Publishing AG 2017
Abstract Soils are significant terrestrial carbon
stores yet the mechanisms that stabilize organic
carbon in mineral soil remain poorly constrained.
Here, we identified climate and topographic controls
on soil organic carbon storage along the Catalina
Critical Zone Observatory that spans a significant
range in mean annual temperature ([10 �C) and mean
annual precipitation ([50 cm year-1). Granitic soils
were collected from divergent summit and convergent
footslope positions in desert scrub, pine, and mixed
conifer systems. Physical soil carbon distribution was
quantified using a density and sonication technique to
obtain the ‘‘free,’’ ‘‘occluded,’’ and heavy ‘‘mineral’’
soil carbon pools. We examined bulk soil (\2 mm)
and density fractions using total carbon (%), stable iso-
topic composition (d13C), and radiocarbon analyses
(D14C). Desert scrub soils stored minimal soil carbon
(\1% by weight) that was partitioned to the heavy
mineral pool. Surprisingly, we identified depleted
D14C in the bulk soil (-9 to -66%) and mineral C
fractions (-72 to -90%) from subsurface weathered
granite in the desert system. The transition to the
productive P. pine ecosystem was met with more soil
C ([3%) that partitioned evenly between the free light
and mineral fractions. Soil C in the P. pine system also
reflected the impact of a moderate severity fire in 2002
that led to modern D14C values for bulk soil and
density fractions. The mixed conifer system contained
a greater proportion of passive occluded C in the
subsurface soils. We observed evidence for modern
fire inputs into the surface soils of the mixed conifer
system in combination with buried charcoal and
occluded C associated with historic fire events.
Convergent landscapes contained higher soil carbon
stocks and depleted D14C relative to adjacent diver-
gent landscapes, suggesting a landscape-level mech-
anism that includes the transport, burial, and
preservation of soil carbon downslope. These data
sets provide insights into ecosystem- and hillslope-
scale variations in soil carbon storage across semiarid
to subhumid environments.
Keywords Climate � Critical zone � Desert soil �Forest soil � Soil organic carbon � Topography
Responsible Editor: Jonathan Sanderman.
Electronic supplementary material The online version ofthis article (doi:10.1007/s10533-017-0360-7) contains supple-mentary material, which is available to authorized users.
R. A. Lybrand � C. Rasmussen
Department of Soil, Water and Environmental Science,
University of Arizona, Tucson, AZ 85721, USA
R. A. Lybrand (&)
Department of Crop and Soil Science, Oregon State
University, Corvallis, OR 97331, USA
e-mail: [email protected]
K. Heckman
Northern Research Station, USDA Forest Service,
Houghton, MI 49931, USA
123
Biogeochemistry (2017) 134:261–277
DOI 10.1007/s10533-017-0360-7
Introduction
Soil organic carbon (SOC), the largest and most
dynamic C pool on Earth’s terrestrial surface (2344 Pg
C; Jobbagy and Jackson 2000; Lal 2004), is an integral
part of the global C cycle (Trumbore et al. 1996).
Forests sequester natural and anthropogenic C as
biomass and contribute inputs to belowground soil C
sinks (Pacala et al. 2001; Ekblad et al. 2013).
Temperate forests in the western U.S. serve as
substantial soil C reservoirs (Homann et al. 2007).
For example, conifer forests in Arizona and California
account for *45% of the regional belowground C
budget despite occupying only a quarter of the area’s
land surface (Rasmussen 2006). Western US forests
are susceptible to climate change because warmer
temperatures and variable precipitation lend to the
upward shift of warmer, drier ecosystems and the
simultaneous narrowing of habitats suitable for high
elevation forests (Crimmins et al. 2010). Mechanisms
that preserve soil C in forested systems must be
identified to better understand future trajectories of
belowground C storage (Seager et al. 2007; IPCC
2013). The Sky Island region of the southwestern US
connects large expanses of desert to forested ecosys-
tems (Coronado Planning Partnership 2008), creating
environmental gradients that comprise large precipi-
tation and temperature differences over short distances
(Whittaker and Niering 1965; Lybrand and Rasmussen
2015). Here, we study how mineral soil C stocks and
stabilization mechanisms vary along a bioclimate
gradient in Arizona.
Mechanisms of soil C stabilization encompass a set
of processes that enhance the probability of SOC to
remain in the system, with the stability of soil organic
matter (SOM) defined as ‘‘a situation where either
SOM molecular composition or concentration or both
remain constant for extended periods of time (Berhe
and Kleber 2013).’’ Selective preservation, spatial
accessibility, and organo-mineral interactions are
mechanisms that stabilize organic C and regulate
residence time in mineral soils (Swanston et al. 2005;
Von Lutzow et al. 2008; Torn et al. 2009; Wagai et al.
2009). Free light, occluded, and mineral soil C pools
preserved by these mechanisms show marked differ-
ences in organic C content, chemical composition,
isotopic abundance, and residence time (Golchin et al.
1994; Sohi et al. 2001; Swanston et al. 2002; Schrumpf
et al. 2013). The ‘‘free’’ or ‘‘light’’ fraction (f-LF) is
the unprotected, inter-aggregate C pool that is plant-
like in nature, and exhibits the shortest mean residence
times in soils (Golchin et al. 1994; Torn et al. 2009).
The ‘‘occluded’’ soil C pool (o-LF) comprises the
‘‘light’’ fraction material that is physically occluded
and made spatially inaccessible to microbes and
enzymes through aggregation or other mechanisms
(Von Lutzow et al. 2008). The ‘‘heavy’’ mineral
fraction (HF) encompasses mineral-associated C that
varies in age and stability (Swanston et al. 2005).
Bulk soil C stocks vary with climate (Trumbore
et al. 1996; Kane et al. 2005) and land cover type
(Trumbore et al. 1995; Jobbagy and Jackson 2000;
Rasmussen and White 2010). Studies along envi-
ronmental gradients document substantial, non-lin-
ear increases of soil C in wetter, more productive
ecosystems (Landi et al. 2003; Dai and Huang
2006; Homann et al. 2007; Djukic et al. 2010;
Meier and Leuschner 2010) and provide insight into
how soils might respond to climate change
(Dahlgren et al. 1997). The non-linear increases in
soil C have been attributed to interactions among
precipitation, temperature, vegetation composition,
and decomposition rates along elevation transects in
the Austrian Limestone Alps (Djukic et al. 2010),
the Sierra Nevada, USA (Dahlgren et al. 1997;
Rasmussen et al. 2007, 2010), the Solling Moun-
tains and Thuringia Basin of Central Germany
(Meier and Leuschner 2010), and in six geograph-
ical regions of China (Dai and Huang 2006). The
Sky Islands of the southwestern US exhibit similar
non-linear increases in soil C stocks that reflect
greater moisture availability and higher rates of net
primary productivity (Whittaker and Niering 1975).
Soil C stocks along such gradients also vary with
landscape position (Lybrand and Rasmussen 2015),
indicating that variability in soil C is driven by
local topography and non-linear increases with
climate.
Landscape position contributes to complex catch-
ment-scale variation in soil C storage when overlaid
with regional bioclimatic controls on soil C stocks
(Schimel et al. 1985; Hook and Burke 2000; Rosen-
bloom et al. 2006; Berhe et al. 2008; Webster et al.
2008; Torn et al. 2009; Hancock et al. 2010). Soil C
concentrations can be *2–3 times higher in conver-
gent, depositional footslope landscapes compared to
adjacent midslope and divergent summit positions as
262 Biogeochemistry (2017) 134:261–277
123
observed in catena studies from the western US and
Northern Territory, Australia (Schimel et al. 1985;
Yoo et al. 2006; Berhe et al. 2008, 2012; Hancock
et al. 2010). The variability of soil C with terrain is
associated with the downslope movement of water,
clay, organic matter, dissolved organic carbon (DOC),
and nutrients (Birkeland 1999). This re-distribution of
water and soil materials contributes to hillslope scale
differences in vegetation composition, biomass accre-
tion, and C inputs to the subsurface (Gessler et al.
1995; Nicolau et al. 1996). A process-response model
demonstrated a high degree of variability in soil C
distribution with terrain, where large ([50%) percent-
ages of C were predicted in deep, lowland soils as a
result of downslope transport and burial (Rosenbloom
et al. 2006). Measured and modeled observations of
soil C storage indicate that preservation mechanisms
and turnover times require examination across three-
dimensional space.
The nature of soil organic matter in fire-prone
forests of the western US reflects intricate soil-
landscape relationships (e.g., Gonzalez-Perez et al.
2004; Berhe et al. 2012; Doetterl et al. 2016). A
review of SOC cycling in dynamic landscapes
describes eroding hillslopes as positions that com-
prise young, labile pools of SOC with limited
subsurface accumulation in soils where fresh SOC
inputs result from net primary production (Doetterl
et al. 2016). Conversely, soils in depositional
positions contain large subsurface C stocks that
represent old, stable pools of SOC where burial
plays a dominant role in C storage processes. The
amount of SOC buried in a depositional landscape
coincides with environmental site conditions as well
as the rate and timing of burial whereas the
composition of the buried SOC can vary (Berhe
et al. 2012), ranging from C-rich surface soils, to
C-depleted subsurface materials, to the integration
of charred materials following wildfire (Doetterl
et al. 2016). A knowledge gap exists in understand-
ing the relative partitioning of free, occluded, and
mineral-associated SOC pools to erosional and
depositional landscape positions, particularly in
fire-impacted systems.
The Sky Island region of the southwestern US has
experienced wildfire as an active, ecosystem altering
process on decadal to centennial timescales (Baisan
and Swetnam 1990; Farris et al. 2010). Patterns of fire
frequency and extent are driven by moisture
availability, topography, and vegetation characteris-
tics of a given landscape (vanWagtendonk et al. 1998;
Stephens et al. 2004; Iniguez et al. 2008, 2016). Tree-
ring reconstructions for the Rincon Mountain Wilder-
ness area in Arizona indicate that large surface fires
([200 ha) burned forested landscapes every decade
over the last several hundred years (Baisan and
Swetnam 1990). We predict that both modern and
historic fire events contribute to the complex nature of
soil C pools in the Sky Island region, particularly in the
productive conifer systems where post-fire erosion and
burial also regulate soil C preservation (Carroll et al.
2007). Quantifying the variation in the partitioning and
storage of SOC along climate and topographic gradi-
ents requires recognition of modern and historic fire
events that have and continue to impact ecosystems.
The objective of this study was to quantify how soil
C stocks vary with depth, topography, and climate in
the transition from desert scrub to fire-prone forests in
the Catalina Critical Zone Observatory (C-CZO) of
southern Arizona, USA. We sampled surface and
subsurface mineral soils from two catena end mem-
bers in desert scrub, pine, and mixed conifer systems
to address the combined impacts of climate, vegeta-
tion, and landscape position on soil C stabilization.
Previous work suggested that footslope positions
concentrate twice as much soil C as adjacent summit
landscapes in these environments (Lybrand and Ras-
mussen 2015). However, the residence time of this C is
unknown, as are the dominant preservation mecha-
nisms in each ecosystem. The impact of historic and
modern fire regimes on the partitioning and storage of
SOC is also poorly understood in the Sky Island
region. We measured organic C and nitrogen (N),
d13C, and D14C for bulk soil (\2 mm) and soil C
fractions obtained through density separation. We
hypothesized that climate exhibits the strongest con-
trol on physical C distribution andD14C in the Catalina
CZO. We expected soil C in the dry, hot desert soils to
stabilize as organo-mineral complexes with short
mean residence times, whereas fire-prone conifer
systems would contain stable soil C preserved as
occluded char material. We further hypothesized that
landscape position would act as a local control on soil
C distribution with greater quantities of stable, old
carbon in water-gathering convergent areas; an obser-
vation predicted to increase proportionally with
moisture availability along the environmental
gradient.
Biogeochemistry (2017) 134:261–277 263
123
Methods
Study site
The Catalina Critical Zone Observatory in Arizona
encompasses an environmental gradient spanning
mean annual temperatures of 21–9 �C and precipita-
tion ranges of 250–950 mm year-1. (Figure 1a).
Estimates of net primary productivity range from 92
to 140 g m-2 year-1 in desert ecosystems compared
to 1050–1150 g m-2 year-1 in high-elevation conifer
forests (Whittaker and Niering 1975). Vegetation
types include desert scrub (Carnegiea gigantean,
Fouquieria splendens, Ferocactus wislizeni, Agave
schotti, and Agave palmeri) at 800 m above sea level
(m a.s.l.), to conifer forest dominated by Pinus
Ponderosa at *2250 m a.s.l., to mixed conifer forest
comprised of Pseudotsuga menziesii, Pinus pon-
derosa, and Abies concolor at 2500 m a.s.l. (Table 1)
(Whittaker and Niering 1965; Whittaker et al. 1968).
Desert scrub, P. pine, and mixed conifer sites were
selected from north-facing hillslopes on granitic
parent materials of similar mineralogical composition
(Lybrand and Rasmussen 2014). The Monitoring
Trends in Burn Severity (MTBS) database was used
to examine the modern fire history of our sites by
identifying fires that were greater than 1000 acres in
size and occurred from 1984 to the present (mtbs.gov;
Eidenshink et al. 2007). The desert scrub site showed
no recent wildfire activity whereas the P. pine site
burned with moderate severity during the Bullock Fire
in 2002 and the mixed conifer site experienced a low
severity burn from the Aspen Fire in 2003.
Aridity class—defined as the ratio of mean annual
precipitation (MAP) to potential evapotranspiration
(PET)— was extracted for each ecosystem from the
PRISM climate dataset (prism.oregonstate.edu; Daly
et al. 2015) and sites classified as water-limited, MAP/
PET\1, or energy-limited, MAP/PET[1. Potential
evapotranspiration was calculated following Thornth-
waite and Mather (1957). The desert scrub ecosystem
had a MAP/PET of 0.5 compared to 1.3 and 1.5 in the
P. pine and mixed conifer ecosystems, respectively
(Table 1; Lybrand and Rasmussen 2015). The desert
scrub soils have an aridic soil moisture regime and
thermic soil temperature regime that transitions to an
ustic soil moisture regime and mesic soil temperature
regime in the P. pine and conifer sites (Soil Survey
Staff 2010).
Sampling scheme
We sampled pedons from two divergent and two
convergent landscape positions within each ecosys-
tem, for a total of twelve pedons (Fig. 1b). Soils from
Fig. 1 a A schematic of the field sites along the Santa Catalina
environmental gradient where soils were collected from
b divergent-convergent landscape positions along a hillslope
within each field site (example in this figure is from the mixed
conifer site). c Surface and subsurface soils were collected for
density separations and radiocarbon analyses
264 Biogeochemistry (2017) 134:261–277
123
water-shedding, divergent summits and adjacent
water-gathering, convergent footslopes were sampled
from transects perpendicular to the slope to minimize
differences in colluvial deposition (Birkeland 1999).
Soils were collected by genetic horizon to the depth of
refusal, air-dried, sieved to\2 mm to obtain the fine-
earth fraction, and prepared for total C and N analysis.
A subset of surface (0–10 cm) and subsurface (30-
40 cm) samples from one divergent-convergent pair in
each ecosystem were used to examine soil C parti-
tioning in density fractions (Fig. 1c).
Density fractions separation
Density separation and sonication were used to obtain
three operationally defined pools of soil C: the f-LF, or
non-mineral associated C isolated by flotation in a
dense liquid; the o-LF, or soil C separated by flotation
in a dense liquid after sonication that includes C either
occluded within aggregates and/or coated with min-
erals; and HF, the remaining soil C that is associated
directly with mineral surfaces and not isolated by
density flotation/sonication (Fig. A1) (Golchin et al.
1994; Sohi et al. 2001; Rasmussen et al. 2005). In
short, 30 g of bulk soil (\2 mm) was suspended for
24 h in sodium polytungstate adjusted to a density of
1.65 g cm-3. The f-LF separated from the sample
following centrifugation was collected with the
supernatant. The remaining bulk soil was dispersed
using ultrasonic energy at 1500 J g-1 soil with a
Branson 450 Sonifier calibrated with DI water (North
1976). The sonication step separated the o-LF follow-
ing the steps for the f-LF. The remaining heavy
material was the HF. After release from the bulk soil,
each fraction was rinsed with aliquots of 0.01 M
CaCl2 and DI water over 0.8 lm filters, and dried at
40 �C.
Additional soil C data
We compiled percent C for[50 surface soils in the
Santa Catalina Mountains and nearby Rincon Moun-
tains to provide context for our work and to expand the
scope of the sample set (Fig. 2). Total organic C
percentages were extracted from published data for 19
surface soil samples from the north and south sides of
the SCM (Whittaker et al. 1968), 19 surface soil
samples from the south side of the SCM (Galioto
1985), 12 surface soil samples from the Rincon
Mountains that are adjacent and geologically similar
to the SCM (Rasmussen 2008), and 3 surface soil
samples from a mixed conifer forest in the SCM
(Heckman et al. 2014). To expand our sample size, we
integrated additionalD14C data into our interpretations
(e.g., Figure 4d) including 8 bulk soil radiocarbon data
points for granitic soils in the Rincon Mountains
Table 1 Field site characteristics and soil pH data for the
desert scrub, P. pine, and mixed conifer soil pedons in the
Catalina Critical Zone Observatory. Field site characteristics
include mean annual temperature (MAT), mean annual
precipitation (MAP), and major overstory vegetation type. Soil
pH data are reported for convergent and divergent landscape
positions as depth-weighted averages ±2r. Field site charac-
teristics and soil pH data were published previously for the
desert scrub and mixed conifer sites (Lybrand and Rasmussen
2014, 2015)
Field Site Landscape
position
Elevation
(m)
MAT
(�C)MAP
(cm)
MAP/
PET
ratio
1:1 Soil:
water
pH ± 2r
1:1 Soil:
1 M KCl
pH ± 2r
Dominant vegetation type
Desert
Scrub
Convergent 1092 18 45 0.53 6.2 ± 0.38 5.0 ± 0.94 Saguaro (Carnegiea gigantean),
Ocotillo (Fouquieria splendens),
Acacia, Arizona Barrel Cactus
(Ferocactus wislizeni), and Agave
(Agave schotti, Agave palmeri)
Desert
Scrub
Divergent 6.2 ± 0.00 4.8 ± 0.10
Ponderosa
Pine
Convergent 2230 10 91 1.26 5.5 ± 0.18 4.0 ± 0.06 Ponderosa pine (Pinus ponderosa)
with sparse Douglas fir (Pseudotsuga
menziesii)Ponderosa
Pine
Divergent 5.6 ± 0.46 4.2 ± 0.06
Mixed
Conifer
Convergent 2408 9.4 95 1.47 5.6 ± 0.70 4.3 ± 0.46 Douglas fir (Pseudotsuga menziesii),
Ponderosa Pine (Pinus ponderosa)
and White fir (Abies concolor)Mixed
Conifer
Divergent 5.7 ± 1.1 4.4 ± 1.2
Biogeochemistry (2017) 134:261–277 265
123
(Rasmussen 2008) and 3 bulk soil and density fraction
data sets from Heckman et al. (2014).
Soil C and N
Percent soil organic C, percent soil organic N, and the
stable isotopic composition of C (d13C) and N (d15N),were measured on all bulk soil (\2 mm) and density
fraction samples at the University of Arizona’s
Environmental Isotope Laboratory. Approximately
3.5 g of sample was ground in a steel canister
containing 3 tungsten carbide ball bearings that was
placed on a ball-mill for 10 min. If too little sample
was recovered for ball-milling (i.e., f-LF and o-LF
fractions), the material was homogenized using a
mortar and pestle. Elemental analyses were performed
on a Finnigan Delta Plus XL (Thermo Fisher Scien-
tific, Bremen, Germany) coupled to an elemental
analyzer (Costech Analytical Technologies Inc.,
Valencia, CA, USA).
Percent C and percent N were reported on an oven-
dry basis and assumed equivalent to organic C and N
because carbonates were not detected in the samples.
Soil C stocks (kg m-2) for the ith horizon were
determined according to:
Ci kg m�2� �
¼ qs � zi � 1� Vrð Þ�C%
100
� ��10 ð1Þ
where qs is the soil bulk density, zi is horizon
thickness, Vr is volumetric rock fragment, and C %
is percent C. The Ci values were calculated for each
horizon. The qs and Vr inputs were estimated follow-
ing the respective methods of Rawls (1983) and Torri
et al. (1994), as outlined previously (Lybrand and
Rasmussen 2015).
Soil C concentration (kg m-3) was calculated as:
Ci kg m�3� �
¼ Ci kg m�2ð Þzi
� �ð2Þ
where Ci represents total horizon C (kg m-2) as
determined in Eq. 1 and zi is total horizon thickness.
The Ci (kg m-3) values were reported on a horizon
basis.
Bulk soil C refers to the % C or Ci (kg m-3)
measured in the fine-earth (\2 mm) fraction for each
soil horizon. Partitioned C refers to the proportion of
total bulk C distributed in the f-LF, o-LF, or HF that
was density separated from a subset of surface
(0–10 cm) or subsurface (30–40 cm) soil samples.
Radiocarbon analyses
Surface and subsurface bulk soil (\2 mm) and density
fraction samples were graphitized for radiocarbon
analysis at the Lawrence Livermore National Labora-
tory. Briefly, each ground sample (*45 mg) was
combusted in a sealed glass test tube containing CuO
and Ag to generate CO2 (g) that was subsequently
reduced to graphite by heating with H2 (g) and a Fe
catalyst as described by Vogel (1987). Radiocarbon
measurements of the resulting graphite targets were
performed by accelerator mass spectrometry at the
Center for Accelerator Mass Spectrometry (CAMS),
Lawrence Livermore National Lab (Davis et al. 1990).
The resulting measurements were corrected for
radioactive decay through normalization to the abso-
lute activity of Oxalic Acid I— an international
radiocarbon standard. Quality control was accom-
plished by analyzing a set of additional well-estab-
lished radiocarbon standards (Oxalic Acid II,
Australian National University sucrose, and TIRI
wood) with the sample unknowns. Radiocarbon con-
tents are reported as fraction modern (Fm) and in units
of D14C for each bulk soil and C fraction analyzed
(Table A1). Analytical errors of the measured samples
averaged ±0.0031 Fm14C or ±3.1% D14C.
Four DOC samples were also collected from
lysimeters installed in convergent landscapes at the
mixed conifer site. Soil water was sampled from the
Fig. 2 Total carbon percentages for surface soils of sites
spanning the Santa Catalina and Rincon environmental gradi-
ents in Arizona. The figure includes data sets adapted from
Galioto (1985), Whittaker et al. (1968), Rasmussen (2008), and
Heckman et al. (2014)
266 Biogeochemistry (2017) 134:261–277
123
soil-saprock boundary where unexpectedly modern
radiocarbon values in bulk soil C and f-LF were
observed. Water samples were filtered through
0.45 lm nylon filters and transported on ice to
Lawrence Livermore National Laboratory where they
were subsequently lyophilized, graphitized, and mea-
sured by AMS as described above.
Statistical analyses
Correlation analyses among environmental variables,
soil properties, and C variables were performed in
JMP V.11.0.0 (SAS Institute Inc., NC, USA)
(Table A2, Table A3). Most variables required natural
log transformation to achieve a normal distribution.
The bulk soil correlation matrix integrated pH 1:1
KCl, C:N ratio, C (kg m-2), and d13C (%) data from
49 soil horizons sampled in our study across the SCM
desert (n = 13), pine (n = 16), and mixed conifer
sites (n = 20). Clay (%), sand (%), C (%), depth, and
MAP/PET were analyzed using 60 samples from 49
soil horizons in our SCM study, 3 SCM mixed conifer
soil horizons (Heckman et al. 2014), and 8 Rincon
Mountain soil horizons (Rasmussen 2008). The bulk
soil D14C variable contained data from 35 samples
including 24 SCM soil horizons in our study, 3 SCM
mixed conifer soil horizons (Heckman et al. 2014),
and 8 Rincon soil horizons (Rasmussen 2008).
The sampling design with only two soil profiles per
landscape position per ecosystem does not allow use
of a full factorial ANOVA for testing differences
among all main effects of ecosystem and landscape
position. To address this limitation, the data were
pooled across the main effects of ecosystem (n = 3)
and landscape position (n = 2). We tested for differ-
ences in % C, C:N, and isotopic composition using
one-way ANOVA.
Results
Soil C and N variation with climate and landscape
position
Surface soil % C increased exponentially with mois-
ture availability along the Santa Catalina and Rincon
environmental gradients (r2 = 0.53, P\ 0.0001)
from\1% in the desert sites to[8% in mixed conifer
forest (Fig. 2). Climate, represented here as MAP/
PET, exhibited the strongest positive correlations with
soil C stocks, soil C:N, silt, and clay (Table A2, A3;
Whittaker et al. 1968); relationships most pronounced
in the convergent water-gathering landscape positions
(Lybrand and Rasmussen 2015). The ratio of conver-
gent to divergent SOC stocks also increased with
MAP/PET, indicating a greater proportion of SOC
stored in convergent landscapes for conifer versus the
desert scrub sites (Fig. 3).
Soil C partitioning, isotopic composition,
and concentration in density fractions
The density fraction data pooled across all sites
indicated significant variation among fractions regard-
less of ecosystem type with respect to partitioned C
(P\ 0.0001), d13C (P = 0.0390), C:N (P\ 0.0001),
and D14C (P = 0.0298) (Fig. 4a–d). The greatest
proportion of total C was distributed to the mineral
fraction with an average of 56 ± 20%, compared to
36 ± 17% in the f-LF, and 9 ± 8% in the o-LF
fractions (Fig. 4a). The HF was generally enriched in
d13C relative to the o-LF and f-LF fractions (Fig. 4b)
and C:N was highest in the o-LF fractions (Fig. 4c).
The density fractions showed large ranges in D14C
content with distributions from-10% to 115% in the
f-LF, -91% to 116% in the o-LF, and -90% to
110% in the HF. The f-LF was most enriched in D14C,
averaging 73 ± 33% which contrasted respective
mean values of 27 ± 64% and 29 ± 55% for the
o-LF and HF (Fig. 4d).
Fig. 3 A convergent to divergent ratio of soil organic carbon
volumetric concentrations (kg m-3) as a function of MAP/PET.
The data points represent pedon averages that are referenced
from Table 2 herein and Fig. 4 in Lybrand and Rasmussen
(2015)
Biogeochemistry (2017) 134:261–277 267
123
Table
2AveragesoilCproperties
±standarddeviation(SD)forfree
light,occluded
light,andheavymineral
density
fractionsas
pooledbyecosystem
.Density
fractionsfrom
desertscrub,P.pine,andmixed
conifer
ecosystem
sweretested
usingOne-
way
ANOVA
withn=
4pedonsanalyzedwithin
each
grouping.Meansfollowed
bythesameletter
within
afieldproperty
foragiven
density
fractionarenotsignificantlydifferent(Tukey’s
test,p\
0.05)
Freelightfraction(f-LF)
Occluded
lightfraction(o-LF)
HeavyMineralFraction(H
F)
Property
Desert
mean±
SD
P.Pine
mean±
SD
Mixed
conifer
mean±
SD
Pvalue
Desert
mean±
SD
P.Pine
mean±
SD
Mixed
conifer
mean±
SD
P-
value
Desert
mean±
SD
P.Pine
mean±
SD
Mixed
Conifer
mean±
SD
P-
value
Partitioned
Carbon
(%by
mass)
20±
9.1
A51±
10.2B
37±
17A,B
0.0205
3±
3.6
A7.5
±4.20A
15±
9.8
A0.0702
78±
12A
42±
10.7
B48±
14B
0.0059
d13C(%
)-25.2
±0.13A
-22.8
±0.61B
-24.8
±0.69A
0.0003
-22.8
±0.21A
-24.6
±0.38B
-24.7
±0.48B
.0019
-21.3
±1.24A
-23.7
±0.78B
-24.2
±0.66B
0.0042
D14C(%
)82.6
±22.9A
86.1
±7.70A
53.8
±54.9A
0.3895
101±
20.4A
53.7
±39.2
A-2.53±
76.9
A0.1635
-16.8
±76.0
A50.8
±47.3
A32.0
±38.0
A0.2650
Total
Carbon
(%)
24.2
±1.84A
32.3
±3.31B
35.8
±1.39B
0.0002
8.80±
2.55A
39.4
±1.48B
42.0
±1.72B
\0.0001
0.32±
0.11A
0.53±
0.35A
1.46±
1.03A
0.0664
C:N
Ratio
(mass
ratio)
17.3
±1.56A
34.3
±6.47B
36.1
±4.49B
0.0005
11.8
±1.27A
38.4
±3.18B
41.2
±9.71B
0.0035
8.30±
0.92A
11.5
±3.18A,B
17.0
±3.89B
0.0074
268 Biogeochemistry (2017) 134:261–277
123
Proportion of C distributed in density fractions
by ecosystem and landscape position
Relative partitioning of soil C by density fraction
varied from greater than 70% to the HF of desert scrub
soils to a more even distribution of SOC to the f-LF
and HF in the two conifer systems (Fig. 5a–c). Visual
differences in the density fractions were also noted
among ecosystems (Fig. A1). The desert soils con-
tained finer plant materials in the f-LF and lighter
colored organics in the o-LF; the conifer soils had
coarser plant parts in the f-LF and black, fine-grained
organics in the o-LF. The mixed conifer soils
contained a greater proportion of C in the o-LF
relative to lower elevation sites, particularly in the
convergent positions.
The desert, P. pine, and mixed conifer soils
presented contrasting hillslope scale trends in soil C
distribution (Fig. 5a–d). The desert scrub soils con-
tained the highest proportion of soil C in the HF at both
landscape positions (Fig. 5a). Soil C in the desert sites
showed a greater partitioning to the f-LF in divergent
sites versus adjacent convergent sites and a similar
distribution to the o-LF between the two positions.
Soil C in the P. pine soils was distributed evenly
between the f-LF and HF in the divergent soils
compared to convergent positions where a greater
percentage of soil C partitioned to the f-LF (Fig. 5b).
The partitioning of soil C to the o-LF fractions was the
same between landscape positions in the P. pine sites.
Occluded soil C in convergent sites was two times
greater than divergent sites in the mixed conifer soils.
Fig. 4 Soil carbon properties across the free light (f-LF),
occluded light fraction (o-LF), and heavy mineral fractions (HF)
for the Santa Catalina Mountain sites. Soil carbon data sets
include the following: a partitioned C, b d13C (%), c C:N, and
d D14C (%). Each box plot presents the median as a black
horizontal line. The 25th and 75th% are represented as vertical
boxes. The 10th and 90th % are indicated by error bars
Biogeochemistry (2017) 134:261–277 269
123
Furthermore, divergent soils from the mixed conifer
system contained twice as much soil C in the f-LF
relative to convergent positions where the greatest
percentage of soil C partitioned to the HF (Fig. 5c;
Table A4).
Soil C in density fractions pooled by ecosystem
Significant differences in d13C (%), % C, and C:N
were identified between the desert and conifer sites
when density fractions were grouped and compared
among ecosystems (Table 2). The d13C isotope com-
position in the f-LF, o-LF, and HF significantly varied
by ecosystem and was most enriched in the o-LF and
HF of the desert scrub ecosystem (Table 2). Total % C
exhibited significant variation among ecosystems in
the f-LF and o-LF compared to no significant differ-
ences for the HF. Similarly, C:N ratios were highest in
the f-LF and o-LF fractions of the P. pine and mixed
conifer soils whereas significantly lower values were
observed in corresponding fractions from the desert
scrub soils.
Soil radiocarbon content in desert and conifer
landscapes
The desert soils presented the most depleted bulk soil
D14C of our study in both the divergent and convergent
subsurface soils (*30–40 cm), with respective D14C
values of -9 ± 2.8% and -66 ± 2.7% (Table 3).
The heavy mineral fractions from the subsurface
desert soils also represent two of the most depleted
D14C values along the SCM gradient including
-72 ± 2.6% and -90 ± 3.6% for divergent and
convergent sites, respectively (Table 3; Table A1).
Modern D14C values for the f-LF and o-LF from
surface soils in both landscapes were nearly identical
and the subsurface soils did not contain o-LF fractions
in either position. Both landscapes showed pro-
nounced relative depletions of D14C with depth in
the HF—with decreases of 99% from 5 to 35 cm.
The P. pine bulk soil and density fractions had
modern radiocarbon values that were all greater than
0%, apart from one HF sample with a D14C value of -
6% in the subsurface soil. The D14C values in the P.
pine ecosystem varied by depth and landscape position
(Table 3). Soil C in the o-LF and f-LF from the
divergent soils were more enriched in D14C with
depth, where values increased by *20 and 15%,
Fig. 5 Partitioning of soil organic carbon into free light (f-LF),
occluded light fraction (o-LF), and heavy mineral fractions (HF)
extracted from bulk soils in the a desert scrub, b P. pine, and
c mixed conifer field sites
270 Biogeochemistry (2017) 134:261–277
123
respectively; whereas the HF was characterized by a
relative depletion in D14C that spanned 48% at
0–10 cm to -6% at 30–40 cm. In contrast, D14C
values decreased with depth for all density fractions in
the convergent position (Table 3).
The D14C values for the mixed conifer bulk soil and
density fractions were depleted at depth (Table 3),
except for the convergent profile. Here, D14C
increased from -15% at 30–40 cm to -8.2%at *120 cm which occurred at the soil-saprock
boundary (based on duplicate measures from the
horizon with values of 0.0 and -16.3%; Fig. 6). The
unexpected increase in radiocarbon content at the soil-
saprock boundary in the convergent soil led us to
examine depth profiles of D14C at the mixed conifer
site, in addition to multiple measures of DOC at
subsurface depths (Table 3; Table A1). Dissolved
organic C samples collected from lysimeters in
convergent positions during March 2013 exhibited
positive D14C values of 43.1% at 39 cm and 18.5% at
65 cm (Fig. 6; Table A1). A third DOC sample from a
65 cm depth lysimeter in January 2013 was depleted
with a D14C value of -19.8%.
Discussion
Soil C properties and partitioning in density
fractions from desert and conifer systems
Soil C partitioning and isotopic compositions exhib-
ited significant variation among density fractions that
was most pronounced in the transition from moisture-
limited desert scrub to energy-limited conifer systems
(Fig. 5; Table 3). Density fractions from the desert
scrub soils contained the lowest C:N of all sites, the
least enriched d13C of the free light fractions, and
substantial depletion in D14C of the mineral fractions
with depth (Table 2; Table A1).The majority of soil C
in the desert scrub soils partitioned to the heavy
mineral fractions that also presented the lowest soil
C:N ratios and most depleted D14C of the heavy
fractions in the study (Table 2). Our results demon-
strate that the oldest, most microbially processed soil
C in the desert scrub system is preserved in the heavy
mineral fraction of subsurface soils.
The conifer soils show a greater proportion of soil C
distributed to the free light and occluded fractions
Table 3 Radiocarbon values (D14C) for soil organic carbon density and aggregate fractions separated from soils collected along the
Santa Catalina Mountain environmental gradient in Arizona
D14C (%)
Free light Occluded light Heavy mineral Bulk
Ecosystem Landscape position/depth
Desert scrub
Divergent/0–10 cm 86 ± 3.1 87 ± 3.2 27 ± 3.0 51 ± 3.0
Divergent/30–40 cm 57 ± 4.0 – -72 ± 2.6 -9 ± 2.8
Convergent/0–10 cm 112 ± 3.2 116 ± 3.7 67 ± 3.1 107 ± 3.2
Convergent/30–40 cm 76 ± 3.1 – -90 ± 3.6 -66 ± 2.7
Ponderosa pine
Divergent/0–10 cm 75 ± 3.1 35 ± 3.0 48 ± 3.0 78 ± 3
Divergent/30–40 cm 89 ± 3.3 56 ± 3.1 -6 ± 2.9 66 ± 3.9
Convergent/0–10 cm 93 ± 3.2 107 ± 3.3 110 ± 3.2 112 ± 3.3
Convergent/30–40 cm 87 ± 3.1 16 ± 2.9 52 ± 3.0 86 ± 3.3
Mixed conifer
Divergent/0–10 cm 115 ± 3.3 86 ± 3.1 78 ± 3.1 96 ± 3.7
Divergent/30–40 cm 80 ± 3.9 -90 ± 2.8 -1 ± 3.0 17 ± 3.0
Convergent/0–10 cm 30 ± 3.1 30 ± 3.1 48 ± 3.2 41 ± 3.1
Convergent/30–40 cm -10 ± 3.0 -36 ± 2.9 3 ± 3.0 -15 ± 3.0
Convergent/110–130 cm 44 ± 4.3 -35 ± 3.5 -55 ± 3.3 -8.15 ± 3.2
Biogeochemistry (2017) 134:261–277 271
123
(Fig. 5b, c), where both present nearly identical soil
C:N and d13C values, yet the occluded fractions are
notably less enriched in D14C than the free light
fractions. The mineral fractions in both the P. pine and
mixed conifer sites exhibited lower C:N ratios than the
light fractions that suggested a greater degree of
microbial processing. Low C:N ratios and enriched
d13C values are indicators of decomposition in forest
soils (Natelhoffer and Fry 1988; Golchin et al. 1994)
where mineral fractions from grassland and conifer
systems have undergone a higher degree of microbial
degradation than the lighter fractions (Baisden and
Parfitt 2007; Natelhoffer and Fry 1988). The lower
C:N ratios in the heavy mineral fractions relative to the
light fractions supported the trend of greater decom-
position in heavy fractions; however, little to no
difference in d13C composition was detected among
the density fractions of the forest soils in this work
(Table 2) or in another study of soil C from a P. pine
forest in southern Arizona that reported similar d13Cranges in density fractions (Heckman et al. 2014).
Despite similarities in d13C isotopic composition,
the D14C values of the density fractions reflect
substantial differences in estimated mean residence
times (Table A6), likely due to the increased parti-
tioning of charcoal inputs to occluded fractions as a
result of wildfire in western US forests (Heckman et al.
2014).We observed a greater proportion of charcoal in
the occluded fractions from the conifer sites compared
to low elevation desert soils (Fig. A1). Our findings
suggest that the physical occlusion of soil C in
aggregates preserves SOC from degradation in forest
soils (Wagai et al. 2009), and indicates that wildfire
may play a role in the partitioning of soil C to occluded
fractions in fire prone conifer systems (Heckman et al.
2014).
Preservation of soil C: desert scrub
Our study is one of the first to examine the variation of
soil C partitioning in desert landscapes and to report
depleted D14C in the weathered granitic bedrock of a
hot, water-limited desert ecosystem. We found min-
eral-associated C to be the largest and most stable soil
C pool in the desert scrub sites with estimated mean
residence times on the order of hundreds of years in the
subsurface (Table A6) compared to prior work in the
Sonoran Desert that documented mean residence
times of 20–40 years for soil C in bulk soil and clay
fractions (McClaran et al. 2008). We hypothesize that
the depleted D14C observed from the weathered
bedrock in our study results from the low primary
productivity of the desert scrub ecosystem. Here,
limited productivity reduces the potential for soil C
exchange with depth, particularly the replacement of
old soil C with modern C inputs in the fractures of
weathered bedrock.
Few studies have examined soil C partitioning in
desert systems; however, an investigation of shrub
expansion on soil C distribution found that *88% of
bulk soil C in a composite grassland profile partitioned
to the heavy mineral fraction (Connin et al. 1997),
which is similar to the mean of *70% in the heavy
fraction from our study. In contrast, carbon from the
upper 5 cm of alluvial soil in an arid, hyperthermic
Sonoran Desert system largely partitioned to the free
light fraction, with little to no organic C in occluded or
mineral fractions (Rasmussen and White 2010; White
et al. 2009). Contrasting sampling depths may account
for the difference between the prevalence of free light
soil C in prior studies compared to mineral-associated
soil C in our study. The soils in our systems also
formed on granitic bedrock that may facilitate the
differential storage of soil C in the fractures of
weathered rock in contrast to soils that developed on
alluvial fan surfaces.
Fig. 6 Radiocarbon values of bulk soils and free light (f-LF),
occluded (o-LF), and heavy mineral fractions (HF) from mixed
conifer soils in the Santa Catalina Mountains, AZ
272 Biogeochemistry (2017) 134:261–277
123
Preservation of soil C: P. pine
Soil C partitioning in the P. pine system shifted
towards a greater proportion of free light and occluded
C (Table 2), with the oldest C in the occluded and
mineral fractions (Table 3). The shift in soil C
partitioning in our study corresponds with the transi-
tion to energy-limited ecosystems that exhibit greater
above- and below-ground primary productivity
(Fig. 1; Whittaker and Niering 1975). Here, larger
proportions of free light and occluded C likely
originate from litter and treefall C inputs in the pine
ecosystem. Greater C concentrations in the mineral
fraction might result from both greater C inputs to the
conifer systems (Fig. 1) and the finer-textured soils in
the P. pine site that facilitate a greater degree of
organo-mineral interactions compared to coarser-
grained desert soils (Lybrand and Rasmussen 2015).
Interestingly, modern D14C values unexpectedly dom-
inated all of the soil C fractions from the P. pine site
(Table 3; Table A1). The young mineral C indicates
an active exchange of mineral-associated C with new
inputs, suggesting a dynamic pool of C pool.
We hypothesize that the soil D14C is dominated by
modern C inputs from the Bullock fire in 2002
(mtbs.gov; Eidenshink et al. 2007). We assume that
the moderate severity burn generated charcoal inputs
comprised of modern C that originated from forest
floor and understory vegetation instead of downed
woody debris that would be complicated by the
inbuilt age of the charcoal in a high severity, stand-
replacing fire (Gavin 2001). Modern, post-fire C
likely entered the P. pine system as a large influx of
particulate, soluble, and charred C, as confirmed
visually by the presence of a charcoal-rich organic
surface horizon and char staining throughout the soil
profile (Fig. A2a). The occluded fraction also
appeared to be largely dominated by charred organic
matter based on visual observations that are consis-
tent with results from other Arizona conifer systems
where occluded fractions were enriched in charred
materials relative to bulk soils (Fig A1; Heckman
et al. 2014). The relative proportion of total C
partitioned to the occluded fraction was greater in the
P. pine site relative to the desert scrub soils.
However, the occluded C in the P. pine site did not
contain the oldest C in the system as predicted, likely
due to the pulse of modern C from the fire into an
otherwise stable SOC pool.
Preservation of soil C: mixed conifer
Soil C in the mixed conifer system was largely
partitioned to the free light and mineral fractions
(Fig. 5c). Soil C in the density fractions of the mixed
conifer landscapes exhibited modern D14C values in
the surface soils that we attribute to inputs from the low
intensity burn during theAspen Fire in 2003. However,
we also observed a substantial shift towards greater
occluded and mineral C in the mixed conifer conver-
gent landscapes that suggests a local hillslope scale
control on soil C partitioning (Fig. A2b). On average,
the mixed conifer soil C exhibited the most depleted
D14C in the subsurface occluded fractions (Fig. 6;
Table 3).We hypothesize that the occluded subsurface
C reflects inputs from earlier wildfire events that were
preserved through post-fire burial processes (Fig. A2b)
as confirmed with the radiocarbon analysis of a
charcoal fragment sampled at 30–40 cm depth
(Fig. 6). This contrasts with the P. pine landscapes
where we only saw evidence for modern inputs from a
recent fire. We also did not observe a thick charcoal-
rich horizon in the mixed conifer site as we did for the
P. pine system (Fig. A2a, b). We speculate that the
absence of the charcoal layer may arise from: a lower
intensity burn at the mixed conifer site (mtbs.gov;
Eidenshink et al. 2007); differences in fuel loads where
long needles of P. pine forests burn quicker than shorter
needles in mixed conifer forests (Iniguez et al. 2008);
or that rapid post-fire erosion in the mixed conifer
forest redistributed the charred layer into materials
deposited in downslope positions.
The occluded fractions from the mixed conifer
subsurface soils had the longest mean residence times
in both landscape positions (Fig. 6). Occluded soil C
represents a stable soil C pool in other forested
ecosystems where turnover times range from tens to
hundreds of years (Rasmussen et al. 2005; Swanston
et al. 2005; Wagai et al. 2009; McFarlane et al. 2013).
The fine-grained black material observed in the
occluded fraction appears to be largely char
(Fig. A1). Occluded C in other AZ conifer systems
exhibited longer mean residence times than free light
and mineral fractions, comprising 0.5–5 times more
charcoal relative to bulk soil as determined by 13C
NMR analyses (Heckman et al. 2014).
The physical distribution and radiocarbon values of
soil C in the mixed conifer system suggest a landscape
level mechanism that includes the downslope
Biogeochemistry (2017) 134:261–277 273
123
transport and burial of soil organic C (Figs. 6, A2b;
Table 3). The mixed conifer convergent soils con-
tained twice the amount of occluded C in the
subsurface versus divergent soils (Table A4), suggest-
ing greater distribution and/or preservation of
occluded materials in the convergent landscapes. We
predict that charred organics partitioned to the
occluded fraction in this environment following
wildfire activity and that the occluded materials
became buried and preserved in the subsurface via
bioturbation or post-fire hillslope erosional processes
(Sanborn et al. 2006).
The preservation of soil C at the hillslope scale
A greater magnitude of SOC accumulated in conver-
gent landscape positions at higher MAP/PET (Fig. 3).
We attribute this to greater in situ net primary
productivity in conifer systems, greater post-fire ero-
sion rates in conifer forests compared to desert scrub
sites, and greater sediment transport rates in high
elevation conifer systems frombioturbation that results
in proportionally more SOC storage in downslope
positions. The thick convergent zone soils in the mixed
conifer system corresponds to higher rates of colluvial
transport and sediment deposition in the high elevation
systems of the Santa Catalina and Pinaleno ranges
modeled in Pelletier et al. (2013).We postulate that the
SOC preserved in these transported sediments con-
tribute to the higher convergent/divergent SOC ratios
detected in the P. pine and mixed conifer soils (Fig. 3).
Density fractions from the desert scrub and mixed
conifer sites show a greater partitioning of SOC to the
f-LF in divergent sites compared to the HF for
convergent positions (Fig. 5; Table A2). The D14C
data from the desert scrub and mixed conifer sites also
reflect modern inputs to surface divergent soils and the
preservation of older C in subsurface soils of deposi-
tional landscapes (Fig. 6; Table 3).We predict that the
prevalence of passive C in subsurface convergent soils
from the desert system results from burial by physical
erosion processes related to a greater degree of bare
ground exposure compared to high elevation forests.
Percent vegetation cover averaged *30% in the
desert scrub site compared to[85% in the conifer
systems (Fig. A3; unpublished data), suggesting that
the exposed desert landscapes would be susceptible to
downslope erosion during high-intensity rain events.
The P. Pine site reflects a more equal distribution
between f-LF and HF by landscape position with a
slightly greater proportion of SOC to f-LF in conver-
gent soils compared to divergent sites. We hypothesize
that younger soil C in the mixed conifer system
dominates the inputs to the f-LF in divergent soils
whereas convergent landscapes reflect an accumula-
tion of older, occluded or mineral-associated SOC that
is often preserved and buried following post-fire
transport from upslope positions (Figs. 5, 6).
The role of dissolved organic carbon in the mixed
conifer system
The bulk soil modern D14C signal at the soil-saprock
interface of the mixed conifer convergent site could be
explained by the downward transport of modern
dissolved organic C (Fig. 6). Modern DOC signals
of soil water collected from convergent subsurface
soils in the mixed conifer watershed support this
hypothesis (Fig. 6; Table A1). The leaching of DOC
from organic surface horizons into the mineral soil
provides an important source of organic C in soils
(Schiff et al. 1990; Michalzik et al. 2003; Rumpel and
Kogel-Knabner 2011), where DOC retained in mineral
horizons account for *22–25% of soil C stocks in
temperate forests (Neff and Asner 2001; Kalbitz et al.
2005). This DOC may persist in the subsurface on
decadal time scales (Sanderman and Amundson
2008). We speculate that the modern radiocarbon
value at[1 m depth in the mixed conifer site results
from the rapid transport of DOC through the coarse,
granitic soils (\10% clay) that subsequently accumu-
lates at the soil-saprock interface (Fig. 6).
Summary
Our results contribute to a limited body of research on
soil C partitioning in desert scrub ecosystemswherewe
identified mineral-associated C in subsurface soils that
exhibited some of the most depleted D14C in our study
(-90%). Greater primary productivity in the conifer
ecosystems led to a more even partitioning of soil C
between free light and heavy mineral fractions in the
forest soil systems, with the greatest proportion of
occluded C in mixed conifer soils. The P. pine system
reflected the impact of a low-moderate severity burn
from the Bullock fire in 2002 that resulted in modern
D14C values for almost all bulk C and density fraction
274 Biogeochemistry (2017) 134:261–277
123
samples. Surface soils in the mixed conifer system also
reflected inputs from a low severity burn associated
with the Aspen fire in 2003; however, we also saw
evidence for historic wildfire in the form of buried
charcoal and occluded materials from subsurface
convergent soils that were both depleted in D14C.
Dissolved organic carbonwas an important component
of the mixed conifer system with radiocarbon data
suggesting deep vertical transport and accumulation at
the soil-saprock interface in convergent positions.
Landscape position represented a hillslope scale con-
trol on soil C partitioning and D14C variation in the
desert scrub and mixed conifer systems, including a
greater proportion of fast-cycling soil C distributed to
free light fractions in divergent positions compared to
more passive mineral-associated C in convergent
landscapes. We attribute the landscape level controls
on soil C distribution to physical erosion resulting from
limited vegetation cover in desert systems and a greater
degree of post-fire erosion and burial following historic
and modern wildfire events in the mixed conifer
system. Our results confirm the interactive role of
climate, vegetation, and landscape position in preserv-
ing soil C along the Catalina environmental gradient–
from mineral-associated C in the shallow soils and
saprock of the desert system to a mixture of soil C
partitioned to free, occluded, and mineral fractions in
the productive, fire-prone conifer forests.
Acknowledgements This research was funded by NSF EAR-
1123454, NSF EAR/IF-0929850, the national Critical Zone
Observatory program via NSF EAR-0724958 and NSF EAR-
0632516, and Geological Society of America’s Graduate
Student Research Grant. Logistical support and/or data were
provided by the NSF supported Niwot Ridge Long-Term
Ecological Research project and the University of Colorado
Mountain Research Station. Radiocarbon analyses were
supported by the Radiocarbon Collaborative, which is a joint
program of the USDA Forest Service, Lawrence Livermore
National Laboratory, and Michigan Technological University.
The authors also thank Dr. S. Mercer Meding, Katarena Matos,
Molly van Dop, Justine Mayo, Stephanie Castro, Christopher
Clingensmith, Mary Kay Amistadi, Nate Abramson, Dr. Julia
Perdrial, Stephan Hlohowskyj, and Andrew Martinez for
laboratory and field assistance.
References
Baisan CH, Swetnam TW (1990) Fire history on a desert
mountain range; Rincon Mountain Wilderness, Arizona,
USA. Can J For Res 20:1559–1569
Baisden WT, Parfitt RL (2007) Bomb 14C enrichment indicates
decadal C pool in deep soil? Biogeochemistry 85:59–68
Baisden WT, Amundson R, Cook AC, Brenner DL (2002)
Turnover and storage of C and N in five density fractions
from California annual grassland surface soils. Glob Bio-
geochem Cycle 16:1117. doi:10.1029/2001GB001822
Berhe AA, Kleber M (2013) Erosion, deposition, and the per-
sistence of soil organic matter: mechanistic considerations
and problems with terminology. Earth Surf Process Landf
38:908–912. doi:10.1002/esp.3408
Berhe AA, Harden JW, Torn MS, Harte J (2008) Linking soil
organic matter dynamics and erosion-induced terrestrial
carbon sequestration at different landform positions.
J Geophys Res 113:G04039
Berhe AA, Harden JW, Torn MS, Kleber M, Burton SD, Harte J
(2012) Persistence of soil organic matter in eroding versus
depositional landformpositions. J GeophysRes 117:G02019
Birkeland PW (1999) Soils and geomorphology. Oxford
University Press, Oxford, p 448
Carroll EM, Miller WW, Johnson DW, Saito L, Qualls RG,
Walker RF (2007) Spatial analysis of a large magnitude
erosion event following a Sierran wildfire. J Environ Qual
36:1105
Connin SL, Virginia RA, Chamberlain CP (1997) Carbon iso-
topes reveal soil organic matter dynamics following arid
land shrub expansion. Oecologia 110:374–386
Coronado Planning Partnership (2008) State of the Coronado
National Forest: an assessment and recommendations for
the 21st Century. www.skyislandaction.org
Crimmins TM, Crimmins MC, Bertelsen CD (2010) Complex
responses to climate drivers in onset of spring flowering
across a semi-arid elevation gradient. J Ecol
98:1042–1051. doi:10.1111/j.1365-2745.2010.01696.x
Dahlgren RA, Boettinger JL, Huntington GL, Amundson RG
(1997) Soil development along an elevational transect in
the western Sierra Nevada, California. Geoderma
78:207–236
Dai W, Huang Y (2006) Relation of soil organic matter con-
centration to climate and altitude in zonal soils of China.
Catena 65:87–94
Daly C, Smith JI, Olson KV (2015) Mapping atmospheric
moisture climatologies across the conterminous United
States. PLoS ONE 10(10):e0141140
Davis JC, Proctor ID, Southon JR, Caffee MW, Heikkinen DW,
Roberts ML, Moore TL, Turteltaub KW, Nelson DE, Loyd
DH, Vogel JS (1990) LLNL/UC AMS facility and research
program: Proc. 5th International Conference on Accelera-
tor Mass Spectrometry, Paris. Nucl Instruments Methods
Phys Res 269–272
Djukic I, Zehetner F, Tatzber M, Gerzabek MH (2010) Soil
organic-matter stocks and characteristics along an Alpine
elevation gradient. J Plant Nutr Soil Sci 173:30–38
Doetterl S, Berhe AA, Nadeu E, Wang Z, Sommer M, Fiener P
(2016) Erosion, deposition and soil carbon: a review of
process-level controls, experimental tools and models to
address C cycling in dynamic landscapes. Earth Sci Rev
154:102–122
Eidenshink J, Schwind B, Brewer K, Zhu Z-L, Quayle B,
Howard S (2007) A project for monitoring trends in burn
severity. Fire Ecol Spec Issue 3:3–21
Biogeochemistry (2017) 134:261–277 275
123
Ekblad et al (2013) The production and turnover of extrama-
trical mycelium of ectomycorrhizal fungi in forest soils:
role in carbon cycling. Plant Soil 366:1–27. doi:10.1007/
s11104-013-1630-3
Farris CA, Baisan CH, Falk D, Yool SR, Swetnam TW (2010)
Spatial and temporal corrobation of a fire-scar based fire
history in a frequently burned ponderosa pine forest. Ecol
Appl 2010:1598–1614
Galioto TR (1985) The influence of elevation on the humic–
fulvic acid ratio in soils of the Santa Catalina Mountains,
Pima County, Arizona. M.S. thesis, University of Arizona,
Tucson
Gavin DG (2001) Estimation of inbuilt age in radiocarbon ages
of soil charcoal for fire history studies. Radiocarbon
43:27–44
Gessler PE, Moore ID, McKenzie NJ, Ryan PJ (1995) Soil-
landscape modelling and spatial prediction of soil attri-
butes. Int J Geogr Inf Syst 9:421–432
Golchin A, Oades JM, Skjemstad JO, Clarke P (1994) Study of
free and occluded particulate organic matter in soils by
solid-state 13C CP/MAS NMR spectroscopy and scanning
electron microscopy. Aust J Soil Res 32:285–309
Gonzalez-Perez JA, Gonzalez-Vila FJ, Almendros G, Knicker H
(2004) The effect of fire on soil organic matter- a review.
Environ Int 30:855–870. doi:10.1016/j.envint.2004.02.003
Hancock GR, Murphy D, Evans KG (2010) Hillslope and
catchment scale soil organic carbon concentration: an
assessment of the role of geomorphology and soil erosion
in an undisturbed environment. Geoderma 155:36–45
Heckman K, Throckmorton H, Clingemsith C, Javier Gonzalez
Vila F, Horwath WR, Knicker H, Rasmussen C (2014)
Factors affecting the molecular structure and mean resi-
dence time of occluded organics in a lithosequence of soils
under ponderosa pine. Soil Biol Biochem 77:1–11
Homann PS, Kapchinske JS, Boyce A (2007) Relations of
mineral-soil C and N to climate and texture: regional dif-
ferences within the conterminous USA. Biogeochemistry
85:303–316. doi:10.1007/s10533-007-9139-6
Hook P, Burke I (2000) Biogeochemistry in a shortgrass land-
scape: control by topography, soil texture, and microcli-
mate. Ecology 81:2686–2703
Iniguez JM, Swetnam TW, Yool SR (2008) Topography
affected landscape fire history patterns in southern Ari-
zona, USA. For Ecol Manag 256:295–303
Iniguez JM, Swetnam TW, Baisan CH (2016) Fire history and
moisture influences on historical forest age structure in the
sky islands of southern Arizona, USA. J Biogeogr
43:85–95
Intergovernmental Panel on Climate Change (IPCC) (2013) The
physical science basis. Contribution of Working group I to
the fifth assessment report of the intergovernmental panel
on climate change. Cambridge University Press, Cam-
bridge, UK
Jobbagy EG, Jackson RB (2000) The vertical distribution of soil
organic carbon and its relation to climate and vegetation.
Ecol Appl 10:423–436
Kalbitz K, Schwesig D, Rethemeyer J, Matzner E (2005) Sta-
bilization of dissolved organic matter by sorption to the
mineral soil. Soil Biol Biochem 37:1319–1331
Kane ES, Valentine DW, Schuur EAG, Dutta K (2005) Soil
carbon stabilization along climate and stand productivity
gradients in black spruce forests of interior Alaska. Can J
For Res 35:2118–2129
Lal R (2004) Carbon sequestration in dryland ecosystems. Envi-
ron Manag 33:528–544. doi:10.1007/s00267-003-9110-9
Landi A, Anderson DW, Mermut AR (2003) Organic carbon
storage and stable isotope composition of soils along a
grassland to forest environmental gradient in Saskatch-
ewan. Can J Soil Sci 83:405–414
Lybrand RA, Rasmussen C (2014) Linking soil element-mass-
transfer to microscale mineral weathering across a semiarid
environmental gradient. Chem Geol 381:26–39
Lybrand RA, Rasmussen C (2015) Quantifying climate and
landscape position controls on soil development in semi-
arid ecosystems. Soil Sci Soc Am J 79:104–116
McClaran MP, Moore-Kucera J, Martens DA, van Haren J,
Marsh SE (2008) Soil carbon and nitrogen in relation to
shrub size and death in a semi-arid grassland. Geoderma
145:60–68
McFarlane KJ, Torn MS, Hanson PJ, Porras RC, Swanston CW,
Callaham MA, Guilderson TP (2013) Comparison of soil
organic matter dynamics at five temperate deciduous for-
ests with physical fractionation and radiocarbon measure-
ments. Biogeochemistry 112:457–476
Meier IC, Leuschner C (2010) Variation of soil and biomass
carbon pools in beech forests across a precipitation gradi-
ent. Glob Chang Biol 16:1035–1045
Michalzik B, Tipping E, Mulder J, Gallardo Lancho JF, Matzner
E, Bryant CL, Clarke N, Lofts S, Vicente Esteban MA
(2003) Modelling the production and transport of dissolved
organic carbon in forest soils. Biogeochemistry
66:241–264
Natelhoffer KJ, Fry B (1988) Controls on natural nitrogen-15
and carbon-13 abundances in forest soil organic matter.
Soil Sci Soc Am J 52:1633–1640
Neff JC, Asner GP (2001) Dissolved organic carbon in terres-
trial ecosystems: synthesis and a model. Ecosystems
4:29–48
Nicolau JM, Sole-Benet A, Puigdefabregas J, Gutierrez L
(1996) Effects of soil and vegetation on runoff along a
catena in semi-arid Spain. Geomorphology 14:297–309
North PF (1976) Towards an absolute measurement of soil
structural stability using ultrasound. J Soil Sci 27:451–459
Pacala SW et al (2001) Consistent land- and atmosphere-based
U.S. carbon sink estimates. Science 292:2316. doi:10.
1126/science.1057320
Pelletier JD et al (2013) Coevolution of nonlinear trends in
vegetation, soils, and topography with elevation and slope
aspect: a case study in the sky islands of southern Arizona.
J Geophys Res 118:741–758. doi:10.1002/jgrf.20046
Rasmussen C (2006) Distribution of soil organic and inorganic
carbon pools by biome and soil taxa in Arizona. Soil Sci
Soc Am J 70:256–265
Rasmussen C (2008) Mass balance of carbon cycling and min-
eral weathering across a semiarid environmental gradient.
Geochim Cosmochim Acta 72:A778
Rasmussen C, White DA (2010) Vegetation effects on soil
organic carbon quality in an arid hyperthermic ecosystem.
Soil Sci 175:438–446
Rasmussen C, Torn MS, Southard RJ (2005) mineral assem-
blage and aggregates control carbon dynamics in a Cali-
fornia conifer forest. Soil Sci Soc Am J 69:1711
276 Biogeochemistry (2017) 134:261–277
123
Rasmussen C, Matsuyama N, Dahlgren RA, Southard RL,
Brauer N (2007) Soil genesis and mineral transformation
across an environmental gradient on andesitic lahar. Soil
Sci Soc Am J 71:225
Rasmussen C, Dahlgren RA, Southard RJ (2010) Basalt
weathering and pedogenesis across an environmental gra-
dient in the southern Cascade Range, California, USA.
Geoderma 154:473–485
Rawls WJ (1983) Estimating soil bulk density from particle size
analysis and organic matter content. Soil Sci 135:123–125.
doi:10.1097/00010694-198302000-00007
Rosenbloom NA, Harden JW, Neff JC, Schimel DS (2006)
Geomorphic control of landscape carbon accumulation.
J Geophys Res 111:G01004
Rumpel C, Kogel-Knabner I (2011) Deep soil organic matter—a
key but poorly understood component of terrestrial C cycle.
Plant Soil 338:143–158
Sanborn P, Geertsema M, Jull AJT, Hawkes B (2006) Soil and
sedimentary charcoal evidence for Holocene forest fires in
an inland temperate rainforest, east-central British
Columbia, Canada. Holocene 16:415–427
Sanderman J, Amundson R (2008) A comparative study of
dissolved organic carbon transport and stabilization in
California forest and grassland soils. Biogeochemistry
92:41–59
Schiff SL, Aravena R, Trumbore SE, Dillon PJ (1990) Dissolved
organic carbon cycling in forested watersheds: a carbon
isotope approach. 26:2949–2957
Schimel D, Stillwell MA, Woodmansee RG (1985) Biogeo-
chemistry of C, N, and P in a soil catena of the shortgrass
steppe. Ecology 66:276–282
Schrumpf M, Kaiser K, Guggenberger G, Persson T, Kogel-
Knaber I, Schulze E-D (2013) Storage and stability of
organic carbon in soils as related to depth, occlusion within
aggregates, and attachment to minerals. Biogeosciences
10:1675–1691
Seager R, Ting M, Held I, Kushnir Y, Lu J, Vecchi G, Huang
H-P, Harnik N, Leetmaa A, Lau N-C, Li C, Velez J, Naik N
(2007) Model projections of an imminent transition to a
more arid climate in southwestern North America. Science
316:1181–1184
Sohi SP, Mahieu N, Arah JRM, Powlson DS, Madari B, Gaunt
JL (2001) A procedure for isolating soil organic matter
fractions suitable for modeling. Soil Sci Soc Am J
65:1121–1128
Soil Survey Staff (2010) Keys to soil taxonomy, 11th edn. U.S.
Government Print Office, Washington, DC
Stephens SL, FinneyMA, Shantz H (2004) Bulk density and fuel
loads of ponderosa pine and white fir forest floors: impacts
of leaf morphology. Northwest Sci 78:93–100
Swanston CW, Caldwell BA, Homann PS, Ganio L, Sollins P
(2002) Carbon dynamics during a long-term incubation of
separate and recombined density fractions from seven
forest soils. Soil Biol Biochem 34:1121–1130
Swanston CW, Torn MS, Hanson PJ, Southon JR, Garten CT,
Hanlon EM, Ganio L (2005) Initial characterization of
processes of soil carbon stabilization using forest stand-
level radiocarbon enrichment. Geoderma 128:52–62
Thornthwaite CW, Mather JR (1957) Instructions and tables for
computing potential evapotranspiration and the water
balance. Publ Climatol 10(3):185–311
Torn MS, Swanston CW, Castanha C, Trumbore SE (2009)
Storage and turnover of organic matter in soil. In: Nicola
Senesi, Baoshan Xing and PMH (eds) Biophys. Process.
Involv. Nat. Nonliving Org. Matter Environ. Syst. John
Wiley & Sons, Inc., pp 219–273
Torri D, Poesen J, Monaci F, Busoni E (1994) Rock fragment
content and fine soil bulk density. Catena 23:65–71. doi:10.
1016/0341-8162(94)90053-1
Trumbore SE, Davidson EA, Barbosa de Camargo P, Nepstad
DC, Martinelli LA (1995) Belowground cycling of carbon
in forests and pastures of Eastern Amazonia. Glob Bio
9:515–528
Trumbore SE,ChadwickOA,AmundsonR (1996)Rapid exchange
between soil carbonandatmospheric carbondioxide drivenby
temperature change. Science 272(80):393–396
van Wagtendonk JW, Benedict JM, Sydoriak WM (1998) Dual
bed characteristics of Sierra Nevada conifers. West J Appl
For 13:73–84
Vogel JS (1987) C-14 background levels in an accelerator mass-
spectrometry system. Radiocarbon 29:323–333
Von Lutzow M, Kogel-Knabner I, Ludwig B, Matzner E, Flessa
H, Ekscmitt K, Guggenberger G, Marschner B, Kalbitz K
(2008) Stabilization mechanisms of organic matter in four
temperate soils: development and application of a con-
ceptual model. J Plant Nutr Soil Sci 171:111–124
Wagai R, Mayer LM, Kitayama K (2009) Nature of the ‘‘oc-
cluded’’ low-density fraction in soil organic matter studies:
a critical review. Soil Sci Plant Nutr 55:13–25
Webster KL, Creed IF, Bourbonniere RA, Beall FD (2008)
Controls on the heterogeneity of soil respiration in a tol-
erant hardwood forest. J Geophys Res 113:G03018
White DA, Welty-Bernard A, Rasmussen C, Schwartz E (2009)
Vegetation controls on soil organic carbon dynamics in an
arid, hyperthermic ecosystem. Geoderma 150:214–223
Whittaker R, NieringW (1965) Vegetation of the Santa Catalina
Mountains, Arizona: a gradient analysis of the south slope.
Ecology 46:429–452
Whittaker R, NieringW (1975) Vegetation of the Santa Catalina
Mountains, Arizona. V. Biomass, production, and diversity
along the elevation gradient. Ecology 56:771–790
Whittaker RH, Buol SW, NieringWA, Havens YH (1968) A soil
and vegetation pattern in the Santa Catalina Mountains,
Arizona. Soil Sci 105:440–450
Yoo K, Amundson R, Heimsath AM, Dietrich WE (2006)
Spatial patterns of soil organic carbon on hillslopes: inte-
grating geomorphic processes and the biological C cycle.
Geoderma 130:47–65
Biogeochemistry (2017) 134:261–277 277
123